Only copy over the direct superclass of our terms from other ontologies

[arteymix]

This makes it much easier to work with the ontology file and in principle shouldn't affect our ability to perform ontology inference.

I have to check first if the inference with Gemma still works properly as we would be lacking the full set of super/sub-class hierarchy and part of/has part relations.

[ppavlidis]

I think I understand what you are saying but you might need to explain why this is a concern? Performance(

We don't use GENO separately except via TGEMO and only need a few terms from it, and are unlikely to use many other ones.

[arteymix]

Yes, this is about performance for both Gemma but also QoL for curators that have to wait 5 minutes for Protégé to load right now.

[arteymix]

I've addressed the issue of inference across ontologies in PavlidisLab/Gemma@4300e77, and I already went ahead with disabling import for TGEMO in tests to speed them up. This will be in the 1.31, so I'll wait until it's out before merging.